In a novel experiment, patients given a placebo, and told it was merely a placebo, showed remarkable improvement in their IBS (irritable bowel syndrome) symptoms. However, more than anything, this study demonstrates the power of positive suggestion, and the researchers overlooked another factor that may affect pain research outcomes and clinical interactions — the Hawthorne effect.
We have previously commented on the powers of placebo in pain management [here], noting that placebos can be effective in more than 30% of patients treated for pain conditions, providing pain-symptom reductions of 50% or greater. At the same time, surveys have observed that more than half of physicians commonly prescribe treatments for certain patients that are expected to have primarily placebo effects [abstract here].
Now, writing in the December 2010 issue of the open-access journal PLoS ONE from the Public Library of Science, researchers at Harvard Medical School examined the “ethically questionable” practice of prescribing placebos to patients who are unaware that they are taking dummy pills [Kaptchuk et al. 2010]. In brief, they found that a group of patients with irritable bowel syndrome (IBS) reported significant symptom relief even when they were told in advance that their medication was fake.
The study was a 2-group, randomized, controlled 3-week trial conducted at a single academic center, involving 80 primarily female (70%) patients, mean age 47, diagnosed with IBS (score ≥150 on the IBS Symptom Severity Scale [IBS-SSS]). Patients were randomized to either twice-daily open-label placebo pills (n=37) — which were presented as “placebo pills made of an inert substance, like sugar pills, that have been shown in clinical studies to produce significant improvement in IBS symptoms through mind-body self-healing processes” — or no-treatment controls (n=43), who had the same quality of interaction with the researchers but received no study therapy.
The primary outcome was IBS Global Improvement Scale (IBS-GIS). This is a 7-category question that asks participants: “Compared to the way you felt before you entered the study, have your IBS symptoms over the past 7 days been: 1) Substantially Worse, 2) Moderately Worse, 3) Slightly Worse, 4) No Change, 5) Slightly Improved, 6) Moderately Improved, or 7) Substantially Improved.” Secondary measures included the IBS Symptom Severity Scale (IBS-SSS), IBS Adequate Relief (IBS-AR), and IBS Quality of Life (IBS-QoL).
The researchers found that the open-label placebo produced significantly higher mean (±SD) global improvement scores (IBS-GIS) at both the 11-day midpoint (5.2±1.0 vs. 4.0±1.1, p<.001) and at the 21-day endpoint (5.0±1.5 vs. 3.9±1.3, p=.002). Significant results were also observed at both time points for reduced symptom severity (IBS-SSS, p=.008 and p=.03) and adequate symptom relief (IBS-AR, p=.02 and p=.03); although, open-label placebo did not show a significant effect for Quality of Life improvement (IBS-QoL) at the 21-day endpoint (p=.08).
In an interview, lead author Ted Kaptchuk, an Associate Professor of Medicine at Harvard Medical School and Beth Israel Deaconess Medical Center, said “Those taking the placebos also doubled their rates of improvement to an almost equivalent level of the effects of the most powerful IBS medications.” The researchers conclude that, “Placebos administered without deception may be an effective treatment for IBS. Further research is warranted in IBS, and perhaps other conditions, to elucidate whether physicians can benefit patients using placebos consistent with informed consent.” The study was funded by the U.S. National Center for Complementary and Alternative Medicine (NCCAM) and the Bernard Osher Foundation.
COMMENTARY: Despite the unique character of this study and seemingly profound results, there are numerous methodological issues that should be questioned, and the outcomes need to be considered in a clinical context. Kaptchuk and colleagues note that irritable bowel syndrome (IBS) is one of the top 10 reasons for seeking primary care, with a world-wide prevalence of approximately 10% to 15%. It is a chronic functional gastrointestinal disorder characterized by abdominal pain and discomfort associated with altered bowel habits, such as constipation or diarrhea.
However, all of the outcome measures in this study relied on subjective self-reports; there was no attempt to more objectively assess changes in symptoms, such as bowel habits. Furthermore, subjects were allowed to continue taking preexisting medications for IBS throughout the study, and higher percentages of subjects in the placebo group were taking antidepressants and other agents, as compared with patients receiving no study-related treatment.
Additionally, the researchers acknowledge that their study was small and, although they conducted a power analysis in advance, the total number of enrolled subjects (80) could provide only strong evidence for large effect sizes in what they termed a “proof-of-principle” pilot study. The ethical propriety of enrolling humans in small, short-term trials with too few subjects to demonstrate potential clinically significant effects in a broader sample of patients has been questioned [Halpern et al. 2002].
Technical note: In their study design and report, the researchers’ use of effect-size calculations based on standardized mean differences [ref. Cohen 1992] does not provide data of practical clinical value and can be erroneously interpreted. For example, they note a large, statistically significant, relative effect size of 0.79 between the placebo and control group at endpoint for the primary outcome, IBS-GIS, and mass media reports touted huge benefits of placebo for IBS. However, looking at the data, there was merely about a 1-point absolute mean difference of from roughly 4 (“No Change”) in the no-treatment group to 5 (“Slightly Improved”) in the placebo group; while the difference was statistically significant, it does not appear to be of much, if any, clinical significance. It is also important to question whether these changes, and the other statistical improvements associated with placebo reported in this study, would persist substantially beyond the brief 3 weeks of the trial. We doubt that they would endure.What is actually most interesting, and clinically relevant, about this novel study is its support for the benefits of positive reinforcement during practitioner-patient interactions. Far from being a deception-free approach, the researchers cleverly manipulated their methodology to “stack the deck,” so to speak, in favor of a placebo effect. To begin, subjects were recruited via advertisements and referrals for "a novel mind-body management study of IBS," and this may have attracted persons particularly interested in alternative medical therapies. Prior to randomization, all subjects were asked if they had heard of the “placebo effect,” and were informed that a placebo pill is an inactive (ie, “inert”) substance like a sugar pill that contains no medication.
However, subjects were further educated on 4 discussion points: a) the placebo effect is powerful, b) the body can automatically respond to taking placebo pills like Pavlov's dogs who salivated when they heard a bell, c) a positive attitude helps but is not necessary, and d) taking the pills faithfully is critical. The researchers concede that, “Our rationale had a positive framing with the aim of optimizing placebo response.” Additionally, Kaptchuk theorized in a published interview that the very ritual of taking pills to treat illness — even fake ones — initiates a brain response that changes the way patients perceive and experience their symptoms. Thus, it would seem that the notion of a completely “inert placebo” may be a non sequitur.
Therefore, the important lesson of this study appears to be: Positive reinforcement offered by practitioners at the time pain medications (or other therapies) are prescribed may have a strong influence on how patients will respond to the treatment. This may seem obvious to many readers, and is reminiscent of a most basic principle of human interaction: ie, the primary influence on people is other people. In everyday clinical settings, the staying-power of such effects might be somewhat ephemeral; however, bolstering responses to genuine therapies by invoking add-on placebo effects in this manner might be highly beneficial.
A second noteworthy aspect of this study, not commented on by the report authors, is the somewhat positive response of subjects in the no-treatment control group. Except for a neutral response on the IBS-GIS measure (noted above), the control subjects did show positive changes to a degree on the other 3 assessments. For example, on the IBS Symptom Severity Scale (IBS-SSS) at 3 weeks control subjects had a mean 46-point reduction compared with a 92-point reduction in placebo subjects; on the adequate symptom relief (IBS-AR) measure, 35% of control subjects compared with 59% of placebo subjects had a positive response.
This suggests that the mere experience of participating in a clinical study exerted a positive, therapeutic influence on patients with IBS even if they received no treatment as part of the study. The phenomenon — generally described as the “Hawthorne effect” and dating back many decades to experiments in social psychology [more information here] — can be an unknown confounder in human research trials. That is, subjects may exhibit improvement in aspects of a condition being studied simply in response to the fact that they know they are being studied, and not in response to any actual treatment. This trial by Kaptchuk et al. is an important demonstration of the “Hawthorne effect” in pain research.
In well-executed randomized, controlled trials, the Hawthorne effect — although present and an underlying contributor to overall placebo and/or treatment effects — would expectedly affect all groups equally and not distort outcomes in terms of the superiority of an experimental treatment compared with a control treatment or placebo. However, this is a major reason that researchers worry about all subjects in a study being treated exactly the same in terms of the interaction with investigators, and this is best managed by having subjects and investigators blinded to group allocation. (Note: neither researchers nor subjects were blinded in this prsent study by Kaptchuk et al.)
An important question for pain management, however, is whether there are aspects of everyday clinical settings and patient-practitioner interactions that can evoke positive responses in patients, completely apart from placebo response or as a consequence of prescribed therapy. That is, to what extent can the mere act of going for pain treatment, and factors associated with the clinical setting and/or human interaction, either positively or negatively influence a patient’s ultimate response to therapy? This seems worthy of consideration.
> Cohen J. A Power Primer. Psychol Bull. 1992;112(1):155-159 [abstract here]
> Halpern SD, Karlawish JH, Berlin JA. The continuing unethical conduct of underpowered clinical trials. JAMA. 2002;17;288(3):358-362 [abstract here].
> Kaptchuk TJ, Friedlander E, Kelley JM, Sanchez MN, Kokkotou E, et al. (2010) Placebos without Deception: A Randomized Controlled Trial in Irritable Bowel Syndrome. PLoS ONE 5(12): e15591 [full article here].